A project that uses street-level cameras on production vehicles and aftermarket telematics systems was able to produce comprehensive, accurate mapping data for horizontal and vertically positioned roadway features, two companies involved in the project announced today. The results of this effort could pave the way for a massive expansion of coverage needed to map the world’s roads in high definition to enable safe autonomy, the companies said in a statement.
An ongoing partnership between Toyota Research Institute-Advanced Development (TRI-AD) and road intelligence company Carmera conducted the latest phase of research. As a follow-up to its camera-based mapping work in Tokyo, the latest phase used commercially available dashboard-mounted cameras to detect key road features with the relative accuracy performance needed for automated driving. The shared objective for both companies is “to achieve significant reduction in costs and massive expansion in geographic scope of autonomy, allowing it to scale to more people in more places.”
Mapping key road features
In the project, Carmera’s machine learning, computer vision and geospatial technologies were used to detect and place key road features – such as lane markings, traffic signals and signs – along Michigan roadways in central Detroit and Ann Arbor. The companies said the results they achieved validated this approach.
“Our work together demonstrates the power of developing systems that are built to take advantage of the ubiquity – and understand the constraints – of automotive technology widely in use today,” said Ro Gupta, CEO of Carmera. “It shows that with the right approach, companies like TRI-AD can start building the data necessary for automated driving, without having to rely on expensive or proprietary third-party hardware. TRI-AD is embracing its leadership position here, and we’re proud to help accelerate this pursuit.”
The project used the same technology that powers Carmera’s Real-Time Events and Change Management engine, which detects, validates and delivers navigational-critical updates to Carmera’s regenerative HD mapping system in minutes rather than month, the company said. The technology can unlock the potential of lower-cost sensors and a vast array of existing vehicle probes to help maintain HD maps everywhere, it added.
“Carmera’s approach has made them a valued partner for our mission,” said Mandali Khalesi, vice president of automated driving strategy and mapping at TRI-AD. “Through the use of production vehicle cameras, we have explored the range of widely available sensor inputs we can draw from, as we look to give automated vehicles around the world a deeper, more accurate understanding of their environments, for a safer driving experience for all.”
Both companies said they look to continue their work and expand their respective technology, partnership, and data platforms.
Carmera, which raised $20 million in funding in 2018, is developing fully regenerative, production HD maps from crowdsourced, camera-only updates. The company supports Level 2 through Level 4 autonomy through its sensor- and format-agnostic approach, and modular map creation and map maintenance offerings. The company is headquartered in New York and Seattle, with offices in San Francisco, London, Detroit, Seoul and Tokyo.
TRI-AD, located in Tokyo, was founded in March 2018 for the purpose of developing innovative products that will enable Toyota’s vision, “Mobility for All.” The group’s mission is to create new technology and advanced safety systems for the world. TRI-AD also oversee’s Toyota’s newly announced Woven City, which will serve as an incubator for smart city design, connected mobility, and robotics technology from Toyota and its partners.